Core-Periphery Structure

A core-periphery structure is a network topology in which a small, densely connected set of nodes (the core) mediates most interactions between a larger, sparsely connected set of nodes (the periphery). The core nodes have high degree and high betweenness centrality; the periphery nodes connect primarily to the core rather than to each other. This topology appears across domains — in the interbank network, in global trade networks, in scientific collaboration networks, and in cortical connectivity — because it optimizes a specific trade: minimizing the path length between any two periphery nodes at the cost of concentrating systemic vulnerability in the core.

The structural implication is double-edged. In normal times, core-periphery networks are efficient: any periphery node can reach any other in at most two hops (periphery-core-periphery). In crisis times, they are fragile: the failure of any core node immediately fragments the network by disconnecting large portions of the periphery from each other. The efficiency-fragility trade is not accidental; it is mathematically necessary. Networks with short average path lengths and high clustering must have either a hierarchical structure or a core-periphery structure, and both concentrate vulnerability in a small number of hubs.

The preferential attachment mechanism that generates scale-free networks also generates core-periphery structure: new nodes attach to high-degree existing nodes, and the high-degree nodes become the core. But not all core-periphery networks are scale-free, and not all scale-free networks have a well-defined core-periphery division. The distinction matters because resilience strategies differ: scale-free networks can be fragmented by targeted hub removal, while core-periphery networks can be stabilized by strengthening the core's internal connectivity — turning a loosely connected core into a tightly connected one reduces the number of single-point-of-failure pathways.